Rev Chem Eng 2017; aop
Muhammad Afzal Raja, Yongsheng Zhao, Xiangping Zhang*, Chunshan Li and Suojiang Zhang
Practices for modeling oil shale pyrolysis and
kinetics
DOI 10.1515/revce-2016-0038
Received September 5, 2016; accepted December 23, 2016
Abstract: Oil shale is one of the largest, relatively unde-
veloped natural fossil fuel resources in the world and
so an important potential source of energy. The organic
matter of oil shale is present as a complex combination
of carbon, hydrogen, sulfur and oxygen named kero-
gen. Pyrolysis-gas chromatography-mass spectroscopy
affords the opportunity to chemically characterize the
main structural skeleton in this kerogen and is a favorable
method to study the structural characteristics of kerogen
at a molecular level. The thermal degradation of oil shale
kerogen is a complex chemical process, accompanied by
the wide variety of products obtained, which poses dif-
ficulties in the determination of the kinetics and mecha-
nism of pyrolysis. Understanding the kinetics of kerogen
decomposition to oil is critical to design a viable retorting
process. Comprehensive kinetic data are also essential for
accurate mathematical modeling of various oil shale pro-
cesses. Classic graphical methods cannot unambiguously
measure and estimate kinetic parameters due to the math-
ematical complexity. Advanced isoconversion methods
would be appropriate for the calculation of the distribu-
tion of activation energies for multiple reactions involved
in the decomposition of complex material such as kerogen
to products. The range of variability in the principal acti-
vation energy is from about 200 to 242 kJ mol − 1, with most
samples being in the middle half of that range, while the
range of frequency factors most likely in the 1012–1016 s − 1
range, with most values within the middle half of that
range. The review presents the complexity of the oil shale
*Corresponding author: Xiangping Zhang, Beijing Key Laboratory
of Ionic Liquids Clean Process, Institute of Process Engineering,
Chinese Academy of Sciences, Beijing 100049, China,
e-mail: xpzhang@ipe.ac.cn
Muhammad Afzal Raja and Yongsheng Zhao: Beijing Key Laboratory
of Ionic Liquids Clean Process, Institute of Process Engineering,
Chinese Academy of Sciences, Beijing 100049, China; and College
of Chemical and Engineering, University of Chinese Academy of
Sciences, Beijing 100049, China
Chunshan Li and Suojiang Zhang: Beijing Key Laboratory of Ionic
Liquids Clean Process, Institute of Process Engineering, Chinese
Academy of Sciences, Beijing 100049, China
pyrolysis mechanism and pyrolysis kinetics along with
the challenges in experimental procedures and modeling
of oil shale pyrolysis kinetics.
Keywords: isoconversion methods; kinetic models; oil
shale; pyrolysis; thermal analysis.
1 Introduction
Energy demand is one of the most important problems
facing the world due to economic and population growth
(Yang et al. 2016b). Demand for crude oil, a cheap source
of energy, currently making up 34.5% (v/v) of the global
energy mix, is expected to grow at 0.8% per annum through
2035 (OPEC 2011, Farouq et al. 2014). The global energy con-
sumption rate made the modern civilization grossly depend-
ent on fossil fuels such as coal, petroleum and natural gas
(Maqsood et al. 2014). Exploration of new reserves and eval-
uation of the potential alternatives have to be considered
together to ensure future energy requirements. To pursue
the alternative energy sources, oil shale, oil sand, shale gas
and coal bed methane have consequently gained worldwide
attention (Jinhu et al. 2014, Youhong et al. 2014). Moreover,
the uncertainty in petroleum prices and finite oil supplies
has motivated many countries to produce and use shale oil
as an alternative to traditional petroleum (Ronghua et  al.
2014, Xiangxin et al. 2014, Yang et al. 2016a).
The early history of pyrolysis kinetics for natural materi-
als showed a huge variability due to improper experimen-
tal design and kinetic analysis methods. Although huge
advances in modeling the free-radical chain reactions associ-
ated with hydrocarbon (HC) were made in the first half of the
20th century, pyrolysis kinetics of coal and oil shale lagged
behind substantially. Over a 20-year period from approxi-
mately 1975 to 1995, rapid advancements in equipment and
software, driven in large part by advances in modern elec-
tronics, provided the methodology to understand the errors
of the past and develop reliable methods that could derive
chemical kinetics capable of extrapolating from geological
time scales of hundreds of millions of years to millisecond
time scales of combustion. Much of this advancement in
understanding the range of variability in kinetic parameters
came from the petroleum exploration geochemistry com-
munity, which desired to extrapolate laboratory-based
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM
2      M.A. Raja et al.: Oil shale pyrolysis and kinetics
kinetic parameters over hundreds of degrees and millions
of years. The International Confederation of Thermal Analy-
sis and Calorimetry (ICTAC) published recommendations
to improve the quality of kinetic parameters derived in the
thermal analysis (TA) community. One of the major recom-
mendations is the need for analyzing multiple thermal his-
tories simultaneously, which had been standard practice in
the petroleum geochemistry community for decades before.
Thus the primary goal of this work is to review the method-
ologies and practices for kinetic parameter determinations
of oil shale. In addition, this work focuses on the challenges
in determination of the oil shale pyrolysis mechanism and
kinetics, and also concludes with the suggestions and
further directions on the way forward.
1.1 Oil shale
Oil shales are natural sedimentary rocks. They consist of
a mineral porous matrix with the porosity filled with oil,
called kerogen, representing 10–65% of the total mass.
The mineral matrix consists of carbonates, quartz and clay
(Martins et  al. 2010). The most important advantages of
oil shale over other solid fossil fuels are high H/C atomic
ratios in the organic matter, which are as high as 1.7 in
some cases (petroleum, 1.9; coal, 0.4–1.0), and a unique
organic matter composition, which allows one to obtain
a wide range of chemical products, including motor fuels
and premium lubricating oils. Figure 1 shows the H/C
atomic ratios in various HC materials and final products.
About 25% of the world oil shale production (or
4 mill. t) is processed into shale oil and combustible gas,
and then upgraded into jet fuel, gasoline, light fuel oil,
bitumen, coke, phenols, liquefied shale gas, wax, lubri-
cating oil and other products (Brendow 2003).
1.2 O
 il shale pyrolysis and retorting
Oil shale is a petroleum source rock that has not under-
gone the natural processes required to convert its organic
matter to oil and gas. Oil shale has been pyrolyzed to
Shale oil,
oil-bearing
Coal sand, and Heavy
Coke and Peat asphalt
anthracite
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5
Figure 1: H/C atomic ratios in various hydrocarbon materials (Strizhakova and Usova 2008).
produce HC oils since the middle of the 19th century (Le
et  al. 2013). The key to produce oil from these resources
is to reduce oil viscosity, and that is best accomplished
by heating these resources up to 500°C, which is known
as the pyrolysis process (Berna et al. 2008). The organic
matter (called kerogen) is less than one third of the shale
weight and cannot be extracted by solvent but is easily
decomposed by pyrolysis (Khraisha 1998). Among the
shale oil recovery technologies, retorting is one of pyro-
lytic processes designed to produce oil and it is considered
as the most effective method for extracting oil from oil
shales (Berna et al. 2008, Haifeng et al. 2015). Generally,
this process is performed by heating the oil shale inside a
vessel called retort in order to promote pyrolysis efficiency
(Vahur and Eric 2013). The composition of the products
formed during pyrolysis depends on the fundamental
chemistry of the organic matter in shale and the reactor
configuration as well (Tiwari and Deo 2012a).
1.3 C
 omposition and chemical structure
of oil shale kerogen
The name “kerogen” was first applied by Crum-Brown in
1912 to denote specifically the insoluble organic matter in
oil shale (Rajeshwar et al. 1979, Sun et al. 2015b). The spec-
ificity of this definition however was lost later. Nowadays
the term “kerogen” is used to characterize solid organic
matter in all sedimentary rocks (Vahur and Eric 2013).
Kerogen in oil shale is a complex heterogeneous mixture of
organic compounds (Doǧan and Uysal 1996). The average
molecular weight is in the order of 3000, and an approxi-
mate empirical formula would be C200H300SN5O11 (Khraisha
1998). Siskin and Katritzky (1995) described that HCs are
rich in the structure of oil shale kerogen and proposed a
kerogen model, shown in Figure 2A, as C647H1017N19O17S4.
Because kerogen has a huge molecular weight of its com-
ponent compounds, it is insoluble in normal organic sol-
vents. Bitumen and/or pre-bitumen, which are known as
the soluble portion of kerogen, may also exist but in rela-
tively lower amounts (Ronghua et al. 2014, Haifeng et al.
2015). Organic matter in oil shale includes the remains of
Premium
petroleum
Petroleum products
Light
Mazut
1.6 1.7 1.8 1.9 2.0
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM
 M.A. Raja et al.: Oil shale pyrolysis and kinetics      3

NH2 H
A OH N
N O O
N
H
N
O OH N
B
H/C = 0.67
N OH 1050 C
O/C = 0.059
O O
OH N N/C = 0.009
N H
H
O NH2 SH

O S
N N N
S O
OH
HO
O
N O
H
N HS OH

Figure 2: (A) Oil shale molecular model C645H1017O17S4N15 proposed by Siskin and Katritzky (1995) and Xiaoping et al. (2015). (B) 3D represen-
tation, using the XMol software, of a type III kerogen structure at the end of catagenesis. Only C atoms (aromatic in white, aliphatic in blue),
O atoms (in dark green) and N atoms (in red) are represented. Published lengths and directions for covalent bonds are used, but energy
minimization calculations are not performed on the resulting structure. However, it is expected to make little difference due to high aroma-
ticity. (Reproduced with permission from Elsevier; Vandenbroucke 2003, Vandenbroucke and Largeau 2007).

algae, spores, pollen, plant cuticle and corky fragments of
herbaceous and woody plants, and other cellular remains
of lacustrine, marine and land plants. Small amounts
of plant resins and waxes also contribute to the organic
matter. Fossil shell and bone fragments composed of
phosphatic and carbonate minerals, although of organic
origin, are excluded from the definition of organic matter
used herein and are considered to be part of the mineral
matrix of the oil shale (Dyni 2003). These materials are
present as a complex combination of carbon, hydrogen,
sulfur and oxygen (Jiang et al. 2007).
The composition and properties of kerogen vary
somewhat from deposit to deposit, but may also vary
within a deposit. Therefore, the properties of kerogen
will be different for the different oil shales. Char is the
organic matter that remains in the spent shale after the
kerogen is paralyzed. Char also has no fixed composition
but depends on the original kerogen and the pyrolysis
conditions (Rick et al. 2009). The general term “kerogen”
cannot represent a defined chemical structure, unlike
other terms such as “protein.” Indeed, the latter corre-
sponds to macromolecules sensu stricto, whereas kerogen
is a complex combination, via numerous bond types, of a
large number of molecular moieties encompassing a wide
range of composition and molecular weight (Katti et  al.
2014). The chemical structure of kerogen is determined
by several main factors; the most important ones are the
composition of initial organic matter and its chemical con-
version during diagenesis. Different mechanisms (when
acting alone) lead to the formation of absolutely different
kerogens with radically different compositions, sets of
the main structural fragments and ways of their linking.
These kerogens have different HC potentials and the main
courses of catagenesis (Bushnev and Burdelnaya 2009).
Based on the fraction of hydrogen, carbon and oxygen
content of the organic matter, kerogens are classified as
type I, type II and type III and are derived from differ-
ent depositional environments (Katti et  al. 2014). Type I
kerogen is composed dominantly of liptinite (alginite) and
has a high atomic H/C ratio. Type III kerogen is derived
from terrestrial plant material and has a low H/C ratio.
Type II kerogen has intermediate properties between type
I and type III and is mainly derived from organic matter
deposited in a marine setting (Brown et al. 2000). A pos-
sible structure obtained for a type III kerogen at the end of
catagenesis is shown in Figure 2B.
1.4 S
 ignificance of modeling the pyrolysis
kinetics of oil shale
Kinetic models are helpful to optimize product yield, effi-
cient product separation, eliminate harmful byproducts
and to find temperature and pressure at which highest
yield of desired product may be achieved. Modeling the
pyrolysis kinetics of oil shale is significant for design-
ing and operating an energy-efficient pyrolysis process
because pyrolysis is a very complicated thermochemical

Brought to you by | University of Colorado - Boulder

Authenticated
Download Date | 4/1/17 10:02 AM
4      M.A. Raja et al.: Oil shale pyrolysis and kinetics
conversion process (Xiaoshu et  al. 2014). The kerogen in
oil shale is such a complicated heterogeneous mixture of
organic compounds that the reported pyrolysis kinetics is
undoubtedly an average for many different reactions that
give the oil product. For this reason, the activation energy
and frequency factor for the oil evolution process are only
“effective” kinetic values for a very complicated reaction
system. They can, however, be used in a mathematical
model to represent the process. Effective activation ener-
gies are extremely useful in modeling a process from an
engineering standpoint. Good kinetic data are essential for
accurate mathematical modeling of various oil shale pro-
cesses (Campbell et al. 1978). Parameters calculated from
a parallel model will provide more reasonable information
for the oil shale retorting process and for recognizing the
chemical structure of oil shale kerogen (Bar et  al. 1988,
Schenk and Horsfield 1993, Shuyuan and C ­ hangtao 2003,
Shuyuan and Changtao 2004, Tiwari and Deo 2012b).
2 E
 xperimental methods to study
pyrolysis kinetics of oil shale
Several investigators have modeled pyrolysis kinetics of
oil shale utilizing data obtained from systems such as
thermogravimetry (TGA), differential analyzers, differen-
tial scanning calorimetry (DSC), Rock-Eval analyzer and
various other types of reactors (Tables 1 and 2). However,
TA is a typical analytical technique to describe the rela-
tionship between physical (or chemical) changes and
temperature. The advantages of TA include convenience,
fast response and fewer samples needed, and thus wide
applications in many fields. With the improvement of
thermo analysis techniques, TA has been increasingly
used as a means of determining kinetic parameters of
fuel pyro­lysis and combustion (Qing et  al. 2007, Sun
et  al. 2015a). Khraisha (1998) described different types
of techniques working in isothermal and non-isothermal
conditions used to study the mechanism of the pyrolytic
process involved in the breakdown of organic matter in
oil shale. For example, TGA-differential thermal analysis
(DTA), Fischer assay and fluidized beds were exercised
to define the kinetic parameters, as well as to find the
quantity and quality of volatile products. Rajeshwar
et al. (1979) concluded that thermal extraction of insolu-
ble kerogen from the oil shale matrix is by far the most
direct and simplest method for processing oil shales. The
TGA-DTA analysis of oil shale samples has been exten-
sively used as a means of determining the characteris-
tics of devolatilization and kinetic parameters (Sun et al.
2014). The pyrolysis of a Yugoslavian oil shale was repre-
sented by three alternative mathematical models, which
were constructed using previously reported TGA- as well
as DSC-derived kinetic parameters, obtained from the
non-isothermal pyrolysis of different samples (Skala
et al. 1989).
Reynolds et  al. (1995) examined three California
shales by programmed temperature micropyrolysis to
determine laboratory pyrolysis kinetics, which measures
HC evolution using a flame ionization detector. The
pyrolysis kinetics were measured by the Pyromat II micro-
pyrolyzer (Lab Instruments, Kenwood, CA, USA) to study
oil shales, petroleum source rocks, coals and bitumen,
in attempts to understand the process of kerogen matu-
ration and HC generation. The three California samples
were analyzed by pyrolysis-mass spectroscopy (MS) uti-
lizing the Pyromat II interfaced with a VG Instruments
Micromass residual gas analyzer (Fisons Instruments,
Danvers, MA, USA). This technique has been described
in detail elsewhere (Burnham et al. 1992b). Stankiewicz
and Kruge (1994) analyzed two samples representing
type II-S kerogen by pyrolysis-gas chromatography-mass
spectroscopy (Py-GC-MS) (CDS 120 pyroprobe, 465, CDS
Analytical, Oxford, PA, USA) and revealed that C1–C5
alkylpyrroles (as well as C1–C4 alkylthiophenes) were
major products in the flash Py-GC-MS analysis of a
Naples Beach kerogen.
2.1 I sothermal and non-isothermal TGA
measurements
Mechanisms of kerogen decomposition have been the
subject of numerous studies (almost all cited articles),
including many different experimental techniques, and
have involved both isothermal and non-isothermal pro-
cedures (Vahur and Eric 2013). Rajeshwar (1981) ana-
lyzed the factors influencing kinetic data such as sample
holder geometry, heating rate and atmosphere using
non-isothermal measurements instead of isothermal to
overcome the significant errors (sample heat-up time,
introduction of correction factor in certain cases and
the large sample size) in the kinetic data. However, the
most effective method for determining the most probable
kinetic mechanism of polymer decomposition (kerogen is
a complicated polymer with a kind of three-dimensional
structure; Shuyuan and Changtao 2003, Tiwari and Deo
2012b) is a combined kinetic analysis of isothermal and
non-isothermal TGA measurements. For example, Camp-
bell et al. (1978) and Torrente and Galán (2001) performed
such a combined analysis.
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM
 Table 1: Critical analysis of kinetic models.
Sr. no.  Sample   Experimental conditions and techniques  
1   Colorado oil   TGA non-isothermal analysis; 2°C min − 1. TGA   k
shale isothermal analysis; 350–424°C. Modified
Fischer assay apparatus was used, oil shale
sample size of <800 μm, sample weight
of 60–80 mg. Inert gas was constantly
flowing through the apparatus during these
experiments.
2   Green River oil   TGA non-isothermal analysis; temperature  
shale range of 25–800°C, ambient inert gas stream
(nitrogen), sample weight of 10–20 mg.
3   Green River oil   The retorts have been operated using shale   LLNL has developed a one-dimensional   The model was used to predict the results for   (Braun et al.
shale with grades ranging in Fischer assay from 42
to 152 L/Mg (10–36 gal/ton); gases such as
air, air and steam, air and N2, air and recycle,
O2 and steam, O2 and CO2, and N2 alone; gas
fluxes ranging from 0.25 to 1.35 mol m − 2 s − 1
(1.2–6.3 scfm/sq ft); shale particles as large
as 30 cm (12 in.); and shale bed void fractions
from 0.49 to 0.29. An attempt was made to
simulate adiabatic conditions in all runs with
heating coils located at the top and bottom
and along the walls of the retort.
4   Green River oil   Aboveground retorting using internal  
shale combustion in a moving packed-bed retort.
Average conditions for the base group of
eight Paraho runs are grade (gal/ton) = 27.0,
raw shale organic carbon (wt%) = 11.8, raw
shale mineral CO2 (wt%) = 17.7, shale velocity
(m/d) = 45.4, bed void fraction = 0.45, cooling
at air distributors (kJ m − 2 s − 1) = 27.0.
Download Date | 4/1/17 10:02 AM
Authenticated
Brought to you by | University of Colorado - Boulder
Proposed models  
/
k   The results show the reliability and
Kerogen →Bitumen → Oil
k1
Oil shale kerogen  →pyrolytic bitumen   The kinetic parameters do not show a
k2
→ oil+gas+semi-coke residue
mathematical model to simulate modified
in situ (MIS) retorting of oil shale.
The one-dimensional mathematical  
model developed at the LLNL to simulate
the detailed chemical reaction kinetics
involved in MIS retorting of oil shale has
been applied to simulate aboveground
retorting in a moving, packed-bed retort
Results/remarks   References
  (Campbell et al.
convenience of non-isothermal kinetic 1978)
experiments in studying oil shale
decomposition reactions. Non-isothermal
experiments more accurately simulate actual
conditions of aboveground and in situ oil shale
retorting. These kinetics are “effective” values
and can only properly be used to describe the
macroscopic oil-production process rather
than the complex microchemistry.
  (Rajeshwar
systematic dependence on heating rate. This is 1981)
compelling proof that the derived parameters
refer to distinct chemical processes rather than
to processes such as product diffusion from the
shale matrix. A key feature of this study is also
a choice of sample holder geometry and sample
mass which facilitates good atmosphere control
and easy escape of product gases.
MIS retorting on a commercial scale, focusing 1984b)
on larger retorts and larger shale particle sizes
than could be investigated experimentally
but the model was not strictly accurate in all
aspects of retorting.
Application of the model to a hypothetical   (Braun et al.
set of commercial-scale conditions helped 1984a)
to identify key design parameters. The
comparisons (data from the Paraho semiworks
retort) illustrated the validity of the model in
calculating retort temperature profiles, oil
yields, off-gas composition and outlet shale
composition. The model calculations pointed
out further resolution, namely,
M.A. Raja et al.: Oil shale pyrolysis and kinetics      5
 Table 1 (continued)
Sr. no.  Sample   Experimental conditions and techniques  
5   Green River oil   The kinetics and product yields calculated  
shale by the model are compared to experimental
results for pyrolysis conditions ranging from
isothermal fluid-bed to high-pressure slow-
heating-rate retorting
6   Yugoslavian oil   TGA and DSC isothermal and non-isothermal   K * 
shale analysis. Isothermal analysis temperatures
673, 708, 748 and 794 K. Non-isothermal
analysis heating rates 2, 5 and 10 K min − 1
Download Date | 4/1/17 10:02 AM
Authenticated
Brought to you by | University of Colorado - Boulder
Proposed models
Governing equations for the model are  
written as 67 first-order, nonlinear,
ordinary equations specifying the rate
of change of each gas, liquid and solid
component in terms of the vaporization or
chemical reactions. This set of equations
is solved simultaneously by Livermore
Solver for Ordinary Differential Equations
on a Cray-1 computer. Most runs took
between 1 and 3 min of CPU time
→ f1B + f2P; f1 + f2 = 1
k1 k2
K *  → f1B1 + f2P1  → f3R1 + f4P2
1 2
K  → f1 B1 + f 2 P 1  → f 3 B 2 + f 4P 2  → f 5 R 1 + f 6 P3
4 6
B → f 7 R2 + f8 P4 → f9 R3 + f10 P5
K* and K: Kerogen and bitumen originally
present in oil shale
P: All volatilized products formed
B: All non-volatilized products
f1 and f2: Stoichiometric coefficients
  Results/remarks   References
(1) the apparent upward displacement of the
combustion zone from the air inlet, (2) the
loss of heat through the wall of the retort and
(3) the possible refluxing of oil mist between
the gas outlet and the combustion zone.
A central feature of this model is that it divides   (Burnham and
the oil into 11 boiling-point fractions in order Braun 1984)
to treat evaporation more rigorously. This
feature also allows cracking of heavy fractions
to lighter fractions, enabling a calculation
of the effect of pyrolysis conditions on the
boiling-point distribution of the product
oil. The model is tested under a wide range
of pyrolysis conditions for which data are
6      M.A. Raja et al.: Oil shale pyrolysis and kinetics
available in the literature. The model does very
well generally, although certain limitations
are demonstrated. It was valid only for Green
River oil shale, but it was expected to be easily
modified for other oil shales.
  Considered retorting process design as   (Skala et al.
the main goal. The advantages of the MSM 1989)
(recognized as a suitable standard procedure
for modeling the kinetics of pyrolysis of
3 different oil shale types) in predicting the
actual heat consumption rates, and in
predicting the pyrolysis effects in both non-
isothermal and isothermal regimes, appear to
be of significance. The results of the modeling
and simulation show that the process can
be simulated using either simple or complex
models.
 Table 1 (continued)
Sr. no.  Sample   Experimental conditions and techniques  
7   Three Turkish   TGA non-isothermal analysis; oil shale  
oil shales samples with three particle sizes of 0.7, 1.3
(Beypazarı, and 2.6 mm, sample weight of 500 mg, flow
Seyitömer and rate of helium was kept constant at 12 L h − 1
Himmetoglu) and temperature range ~27–700°C
8   Jordan shales   TGA isothermal analysis; temperature was  
(EL-Lajjun and varied from 375 to 475°C, atmospheric
Sultani shales) pressure, oil shale sample size of 1.8 × 10 − 4
to 3.55 × 10 − 4 m, sample weight of 0.5 g,
nitrogen flow rate is 1.67 × 10 − 5 m3 s − 1
9   Eight Turkish   TGA and DTG non-isothermal analysis;  
oil shale sample weight of 10 mg, samples particle
size <60 mesh, heating rate of 5°C min − 1,
temperature range of 20–600°C and
50 mL min − 1 gas flow rate (air and nitrogen)
10   Spanish   TGA isothermal analysis; different particle  
oil shale sizes (0.250 × 10 − 3, 0.505 × 10 − 3 and
(Puertollano in 0.925 × 10 − 3 m), three temperature plateaux
Ciudad Real) (703, 688 and 673 K). TGA non-isothermal
analysis; different heating rates (5, 10, 15, 20
and 50 K min − 1) up to 1173 K and for a particle
size of 0.250 × 10 − 3 m
11   Huadian oil   Pyrolysis characteristics studied by TGA,  
shale different particle size diameters, under
nitrogen atmosphere, heating rate of 40°C
min − 1
Download Date | 4/1/17 10:02 AM
Authenticated
Brought to you by | University of Colorado - Boulder
Proposed models   Results/remarks   References
k1
Kerogen  →Bitumen + oil + gas + residue   A first-order overall kinetic expression was   (Doǧan and
k2 found to be adequate for pyrolysis of the three Uysal 1996)
Bitumen → oil + gas + residue shales studied. The extent of pyrolysis was
practically independent of the particle size.
Examination of porosities and surface areas
of raw and spent shales obtained after non-
isothermal pyrolysis up to final temperatures
of 400, 550 and 700°C confirmed the TGA
finding that the higher the final temperature,
the higher the conversion.
Kerogen 
→Bitumen + gas   Kinetic parameters, weight vs. time data   (Khraisha 1998)
Bitumen 
→ Oil + gas + char were obtained at temperatures from 375 to
475°C. The rate of kerogen decomposition
can adequately be described by a simple
mechanism involving first-order kinetics.
Possibly, two consecutive first-order reactions
may take place during the decomposition
process.
k1 k2
Kerogen  →Bitumen  → Oil, gas   Kinetic parameters (both pyrolysis and   (Kök and Pamir
combustion) of oil shale samples were 2000)
determined using five different methods.
Differences between the activation energy
values of the samples were observed. At the
same time different degrees of maturity and
structure of oil shales affected the results.
k1
Oil shale kerogen  →pyrolytic bitumen   The kinetics are “effective” values (useful for   (Torrente and
k2 modeling a complex chemical process) which Galán 2001)
→ oil+gas+semi-coke residue
may only be used to describe the macroscopic
kerogen decomposition process rather than
the microscopic chemistry involved. The
integral method yields lower deviation and
hence provides a better fit of the data.
k1 k2
Kerogen  →Bitumen  → Oil, gas   A comprehensive utilization technology for   (Jiang et al.
Huadian oil shale resources is recommended 2007)
for shale oil production, electricity generation,
oil shale ash processing, economical efficiency
and environmental protection.
M.A. Raja et al.: Oil shale pyrolysis and kinetics      7
 Table 1 (continued)
Sr. no.  Sample   Experimental conditions and techniques  
12   Kukersite oil   Isothermal retorting; 500 mL aluminum retort,     A first-order parallel-consequent kinetic model   (Johannes and
shale the initial material applied was dry powdered
kukersite with kerogen content 32.3%. The
isothermal period was kept for 20 min. The
nominal temperatures applied were chosen
in the region optimal for formation of TB
(thermobitumen) 370–410°C
13   Jordan shales   TGA non-isothermal analysis; oil shale  
(EL-Lajjun) samples size of 50 μm, sample weight of
8–15 g, nitrogen gas is used at 100 mL min − 1,
heating rates of 5, 10, 15 and 20°C min − 1
14   Mahogany   TGA non-isothermal analysis; oil shale sample   Kerogen 
zone of the size of 1.49 × 10 − 4 m, weight of samples
Green River 20–30 mg, elemental analyses of the samples
formation were performed using LECO CHNS-932 and
VTF-900 units, nitrogen environment, seven
different heating rates range from 0.5 to
50°C min − 1.
15   Chinese   TGA/DTA non-isothermal analysis; ambient  
oil shales atmosphere and nitrogen atmosphere (fluxed
(Huadian) by nitrogen with a flow rate of 50 cm3 min − 1,
oil shale sample size of 63 μm, sample weight
of 4 mg, heating rate was kept as 10°C min − 1,
temperature range from ambient to 800°C.
Download Date | 4/1/17 10:02 AM
Authenticated
Brought to you by | University of Colorado - Boulder
Proposed models   Results/remarks   References
Kerogen
k1 k2 was deduced for preliminary description of Zaidentsal
the co-effect of time, nominal temperature 2008)
k4 and heating rate on the yield of products at
TB Voltile
low-temperature pyrolysis of kukersite in
k3 laboratory retorts. The maximum total yield
of TB (thermobitumen) and oil from kerogen
Coke (91%) was obtained at nominal temperature
410°C after 40-min processing.
Kerogen 
→Bitumen + gas   The calculated kinetic parameter results were   (Shabbar et al.
Bitumen 
→ Oil+gas + char compared with the experimental curve of TGA 2011)
analysis and also among themselves. A root
mean square error was calculated for each
method and the integral method model was
8      M.A. Raja et al.: Oil shale pyrolysis and kinetics
found to be the closest to experimental data,
while no relationship was observed between
activation energy and the heating rate.
→Decomposition products   The RMS errors between the experimental and   (Tiwari and Deo
model values for the isoconversion approach 2012b)
produced the lowest RMS values in both rates
and cumulative conversion. Isoconversion
models are in theory “kinetic model” free and
their applicability to the decomposition of a
complex material such as kerogen is excellent.
However, their applicability in reproducing
multistep kinetics has been questioned.
Application of these models to real-life
processes requires extending these models
outside of the experimental data range from
which they were derived.
→Bitumen + gas
Kerogen    The analytical model is successfully validated   (Xiaoshu et al.
Bitumen 
→ Oil+gas + char by both experimental data and the numerical 2014)
model (R2 > 0.99). Some generalizations and
extensions of representative kinetic models in
wide-area applications are presented. There
is a need for such models in fuel and energy
technology.
 Table 2: Summary of graphical methods to determine kinetic parameters.

Sr. no.   Experimental   Modeling techniques/methods   Relative error/remarks   References

technique/method

1   TGA   Arrhenius, Coats Redfern and   The kinetic parameters do not show a systematic dependence on heating rate.   (Rajeshwar 1981)
Non-isothermal Freeman Carroll techniques This is compelling proof that the derived parameters refer to distinct chemical
have been applied processes rather than to processes such as product diffusion from the shale
Differential method developed matrix. A key feature of this study is also a choice of sample holder geometry
by Freeman Carroll and sample mass which facilitates good atmosphere control and easy escape of
product gases.
2   TGA and DSC   Non-isothermal TGA analysis:   Applying the statistical procedure, the slope and intercept of the linear part of   (Skala et al. 1988)
Isothermal and using the Doyle-Gorbachev and the plot determine the activation energy (E) and frequency factor (A).
non-isothermal integral method
3   TGA   Coats Redfern method   The integral method gives the graphical representation to achieve required   (Ahmad and Williams 1998)
Non-isothermal kinetic data.
4   TGA   Arrhenius and Coats Redfern   The decomposition of oil shale kerogen in terms of a first-order reaction appears   (Williams and Ahmad 2000)
Isothermal and to be a suitable assumption since a value of n = 1 offers a best fit regression line
non-isothermal giving a correlation coefficient of 0.999.
5   TGA   Arrhenius and integral method   The assumption of a first-order reaction appears reasonable, because the   (Jaber and Probert 1999)
Non-isothermal regression correlation factor exceeded 0.95.
6   TGA   Flynn’s isoconversional   The maximal relative error of the weight loss for the organic decomposition   (Gersten et al. 2000)
Non-isothermal method, Coats Redfern or stage, calculated by means of the set of kinetic parameters, is 4.9% at
Piloyan-Novikova methods 429.85°C.
7   TGA   Arrhenius and integral method   The assumption of a first-order reaction appears reasonable: the regression   (Jaber and Probert 2000)
Non-isothermal correlation factor exceeded 0.95.
8   Fixed bed retort   A modified first-order kinetic   The obtained kinetic data are modeled using variable heating rate, pyrolysis   (Omar et al. 2010)
equation with variable temperature and variable activation energy principle in a nitrogen sweeping
activation energy is employed medium. Good fit to the obtained experimental data is achieved.
to model the total weight loss
of El-Lajjun oil shale samples
9   TGA   Integral method   Closely simulates the conditions that would be found in commercial scale   (Jaber et al. 1999)
Non-isothermal retorts. The regression correlation factor was > 0.95.
10   Isothermal   Arrhenius method   A step-by-step mathematical model was deduced for description of the co-effect   (Johannes and Zaidentsal 2008)
retorting of time, temperature and heating rate at low temperature (370–410°C)

Download Date | 4/1/17 10:02 AM

Authenticated
Brought to you by | University of Colorado - Boulder
M.A. Raja et al.: Oil shale pyrolysis and kinetics      9
10      M.A. Raja et al.: Oil shale pyrolysis and kinetics
2.2 A
 dvantages of non-isothermal
conditions
In non-isothermal TGA, the solid is submitted to a gradual
increase in temperature (usually linear with time), and in
this case it is possible, in principle, to obtain the kinetic
parameters (activation energy and pre-exponential con-
stant) with a single experiment. Non-isothermal TGA
allows for a wide range of temperatures to be covered,
which is not always possible in isothermal TGA, particu-
larly at high temperatures, since significant solid conver-
sion may occur during temperature stabilization at the
beginning of the experiments (Brown et al. 2000). On the
other hand, one of the problems associated with rising
temperature experiments is that the kinetic parameters
obtained are frequently dependent on some procedural
factors (heating rate, initial amount of solid sample, gas
phase composition, reactor geometry, particle size, solid
porosity). However, this situation also arises in isothermal
studies (Órfão and Martins 2002).
The non-isothermal method has the advantage of
shorter experimental time and eliminates the difficulties
associated with the isothermal method due to the initial
heat-up period. Since a sample may undergo significant
reaction while being raised to the temperature of interest,
the results obtained by the isothermal method are often
questionable (Rajeshwar 1981). The non-isothermal tech-
nique has been preferred by some researchers because of
its advantages over the isothermal method. Some of the
major advantages of non-isothermal technique are the
elimination of errors due to the thermal induction period,
maintaining a constant heating rate is less demanding
compared to retaining a constant temperature environment
during an experiment (especially when exothermic reac-
tions are involved), permits a rapid scan of the whole tem-
perature range of interest and it also simulates conditions
expected in large scale oil shale retorting processes more
closely (Ahmad and Williams 1998, ­Torrente and Galan
2001, Shabbar et al. 2011, Tiwari and Deo 2012b).
3 O
 il shale pyrolysis mechanism
and kinetics
3.1 O
 il shale pyrolysis mechanism
The thermal decomposition of coal, kerogen and most syn-
thetic polymers involves a complicated set of concurrent
and consecutive initiation, propagation and termination
reactions (Burnham and Braun 1999). Oil shale retorting
theories (Hubbard and Robinson 1950, Khraisha 1998,
Xiaoshu et  al. 2014) explain the conversion of kerogen
into shale oil as a two-stage process. Generally, thermal
decomposition of oil shale yields volatile materials in the
form of oil and gas and a solid residue of char (Shabbar
et al. 2011). The pyrolysates of oil shale result mainly from
the rupture of weak chemical bond, probably the rupture
of weak cross-linked bonds, such as C–O and C–S bonds,
and/or the rupture of branched functional groups in the
kerogen long molecular structure. These weak bonds have
low rupture energy which commensurate with pyroly-
sis gaseous products such as H2O, CO2, H2S, H2 and light
HCs. The medium activation energy values are associated
with breakdown of the side chains in β-site of aromatics,
decomposition of normal alkane with large molecular
weight, Diels-Alder cyclization reaction and the rupture of
alicyclic HC (Deepak and Eric 1987, Omar et al. 2010).
Xiaoping et al. (2015) carried out molecular dynam-
ics (MD) simulations using a reactive force field (ReaxFF)
method for Green River oil shale as shown in Figure 3.
The model demonstrates that the thermal decomposition
of the oil shale molecule is initiated with the cleavage of
the oxygen bridge (C−O bond) and the first product is for-
maldehyde (CH2O). The simulation results show that the
C−O bond is weaker than the other bonds, agreeing with
its smaller bond dissociation energy. The ring-opening
position of the aliphatic ring is usually determined by
the stability of free radicals formed in this process. For
aromatic HCs, the long-chain substituents are found to
be easier to leave and the cleavage of C−C bonds leads
Figure 3: Model applied to ReaxFF MD simulations of thermal
decomposition (Xiaoping et al. 2015).
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM

to a series of chain reactions and the formation of small
fragments, such as ethylene and propylene. The bond
cleavages are almost in accordance with the minimum
bonding energy rule. NVT (constant volume and con-
stant temperature) simulations show that the pyrolysis
process progresses in two stages: the decomposition of
kerogen into heavy (C40 +) species and then the generation
of light compounds. Recombination and rearrangements
of different fragments are also observed via MD simula-
tions. The main HC fragments of C10–C20 are regarded as
the component or precursor of diesel oil. The formation
pathways of typical aromatic components are analyzed
by tracking the motion trajectories of relevant structures.
The products obtained from the simulations were similar
to the result of GC-MS analysis, including a great amount
of alkanes, olefins and aromatic compounds (Xiaoping
et  al. 2015). In addition, the main components of the
diesel oil were observed in the simulation process, indi-
cating that the ReaxFF simulations can reproduce the
experimental results under a suitable NVT simulation
temperature and time. Tissot and Welte (1984) classify
kerogens as types I, II and III based on their pyrolytic HC
yield per total organic carbon (TOC). Although Fischer
assay pyrolysis has been the standard for measuring oil
yield from oil shale, Rock-Eval analysis is more com-
monly used in petroleum geochemistry, and the two
methods correlate well after making adjustments for gas
yields. Type I kerogens yield 650–1000 mg HC/g TOC, and
type II kerogens yield 300–650  mg HC/g TOC. Several
investigators tried to describe the pyrolysis mechanism of
type I kerogen (McKee and Lyder 1921, Johnson et al. 1975,
Ziegel and Gorman 1980, Burnham et  al. 1996, Freund
et al. 2007, Burnham 2010, 2015, Le et al. 2013), and many
others analyzed type II kerogen as well (Coburn et  al.
1988, Burnham 1990, Castelli et  al. 1990, Braun et  al.
1991, Burnham et  al. 1992a, Freund et  al. 1993, Behar
et al. 2008). McKee and Lyder (1921) explained the break-
down of oil shale as halving two steps, Eq. (1): the initial
formation of heavy bitumen and the next conversion of
bitumen to oil. The description of their experiments was
formally correct; some had extended the observation to a
sequential chemical reaction mechanism:
Kerogen  → Oil + Gas (1)
→ Bitumen 
Although it is pervasive in the oil shale and petroleum
geochemistry literature, it is simple to show that this
extension is not rigorously correct.
Ziegel and Gorman (1980) showed that not only were
the Braun-Rothman bitumen generation kinetics incon-
sistent with the bitumen concentrations, but also no good
M.A. Raja et al.: Oil shale pyrolysis and kinetics      11
fit was possible with a strictly sequential model. Instead,
they found that the data required an alternate pathway
mechanism of the form
Kerogen 
→ Bitumen
Oil + Gas (2)

It merely states that the reactions forming soluble
products operate in parallel with those generating smaller
volatile products. This can be easily understood in the
context of mechanistic models of kerogen decomposition
and oil generation (Freund et  al. 2007), which use real
chemical species and not hypothetical species that are
really broad mixtures separated by a physical separation
(single-plate distillation) that depends on temperature
and pressure.
Although numerous workers continue to consider
the conversion of kerogen to gas as a sequential reac-
tion involving bitumen intermediate, simple experiments
and kinetic modeling shows that such a simple picture
is not correct. To be sure, the initial breakdown prod-
ucts are mostly higher molecular weight, and they con-
tinue to break down with continued thermal exposure.
However, the species normally defined as bitumen and
oil have overlapping compositions, and the physical split
between the two depends on temperature and pressure.
They are not appropriate lumped species for a chemical
kinetic model. Furthermore, for most coals and kero-
gens, the reactions generating fragments small enough
to vaporize from pendant groups proceed in parallel.
Well-preserved algal kerogens are an exception, and by
analogy with linear synthetic polymers, it appears that
multiple bonds must be broken to form species small
enough to vaporize. This reaction sequence leads to sig-
moidal reaction character. Table 1 presents the analyses
of several investigators regarding thermal breakdown of
oil shale via different experimental procedures and oper-
ating conditions.
3.2 Oil shale pyrolysis kinetics
As with the mechanistic aspects discussed in the previ-
ous section, the results reported by various investigators
on the kinetics of decomposition of oil shale kerogen
reveal a spectrum of widely differing behavior. This trend
is reflected in the postulated decomposition schemes
(Table 1), which range from simple one-step consecu-
tive reactions to complicated branched schemes. The
papers cited in Table 1 direct that the decomposition
process may be viewed as consisting of multiple parallel
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM
12      M.A. Raja et al.: Oil shale pyrolysis and kinetics
reactions with individual activation energies. Shuyuan
and ­Changtao (2003) studied the kinetics of pyrolysis of
oil shale samples from Fushun, Maoming and Huangxian
(People’s Republic of China) using TGA. They concluded
that the dependence of the degree of kerogen conversion
into shale oil and gaseous products on activation energies
has an exponential character. They found that the reac-
tions of pyrolysis with low activation energies result from
the rupture of the weakest chemical bonds. The reactions
of pyrolysis with medium activation energies involve the
rupture of side chains in aromatic compounds (at the β
position), the cleavage of high-molecular-weight alkanes,
cyclization reactions and the cleavage of alicyclic HCs,
which are accompanied by the release of shale oil. The
aromatization of alicyclic compounds, dehydrogeniza-
tion, polymerization and the decomposition of heterocy-
clic compounds belong to the reactions of pyrolysis with
high activation energies.
The temperature is an important determinant param-
eter for activation energy and pre-exponential coefficient,
and accurate estimates of their temperature dependence
are critical to understanding thermal activation pro-
cesses (Jian et al. 2010). There is a wide variation in the
literature for values of activation energy. The type of oil
shale and type of pyrolysis, whether isothermal or non-
isothermal, are influential factors for the discrepancies.
Since a sample may undergo significant reaction while
being raised to the temperature of interest, the results
obtained by the isothermal method are often question-
able. However, this error may be minimized by very rapid
heating to the desired isothermal temperature (Doǧan and
Uysal 1996). Skala and Sokic (1992) developed a kinetic
expression commonly used in the TA of oil shale pyrolysis.
This was derived on the basis of a simple first-order kinetic
equation of kerogen decomposition. The obtained results
show that the largest activation energies were detected by
using isothermal TG, while combined non-isothermal and
isothermal TG gave the smallest values. In all the exam-
ined samples and TG analyses performed, there was an
increase in the activation energy with increased content
of paraffinic structures in the oil shale.
The kinetic parameters do not show a systematic
dependence on heating rate. This is a compelling proof
that the derived parameters refer to distinct chemical
processes rather than to processes such as product dif-
fusion from the shale matrix (Rajeshwar 1981). Although
some attempts (Doǧan and Uysal 1996, Jaber and Probert
1999) have been made to understand the complex nature
of decomposition of kerogen involving numerous organic
compounds, some authors have found it sufficient to con-
sider a global first-order kinetic expression to represent
the overall decomposition rate of kerogen through experi-
ments with small particles. Indeed, an overall kinetic
expression may be quite reliable, especially for most
engineering purposes. But the thermal decomposition
of coal, kerogen and most synthetic polymers involves a
complicated set of concurrent and consecutive initiation,
propagation and termination reactions. Consequently,
essentially none can be described by a first-order reaction.
Therefore, using a first-order analysis method such as
Coats Redfern (Coats and Redfern 1964) on single-heating-
rate data will not result in meaningful kinetic parameters,
and the ICTAC Kinetics Committee (Brown et  al. 2000,
Vyazovkin et al. 2011, 2014) specifically discouraged that
approach. Reporting Coats Redfern A and E values at mul-
tiple heating rates makes no sense whatsoever (Vyazovkin
et al. 2011, 2014).
Mohammad et al. (2011) studied the effect of mineral
matter content on the activation energy of oil shale pyrol-
ysis. Mineral content of oil shale plays a vital role in HC
evolution and pyrolysis reactions. Both activation energy
and pre-exponential factor for oil shale and isolated
kerogen pyrolysis are found to increase with the increase
of heating rate due to heat and mass transfer effects. The
derived values of the apparent activation energy are con-
sistent with a first-order reaction. The mean values of acti-
vation energy are found to be 75 and 95 kJ mol − 1 for raw
oil shale and isolated kerogen pyrolysis, respectively. It is
concluded that the mineral matter in oil shale enhances
catalytic cracking as is evident from the reduced E values
of oil shale compared with those for kerogen. Skala et al.
(1987) reported that the activation energy and frequency
factor of kerogen concentrate prepared by acid deminer-
alization are lower than that of raw oil shale. However,
­Karabakan and Yürüm (1998) found that the activation
energy of pyrolysis reactions of raw oil shale from two
sources was less than that of carbonate-free oil shale. This
indicated that metal cations (M2 +) in the carbonate miner-
als might have acted as catalysts in the pyrolysis reactions.
On the other hand, they found that the activation energy
of pyrolysis reactions of silicate-free oil shale was lower
than the activation energy for both raw and carbonate-free
oil shale. This was explained by the fact that silicate min-
erals acted as inhibitors in pyrolysis reactions.
Burnham (2015) analyzed that a significant portion of
the diversity of A and E values could be attributed to vola-
tility changes over the wide temperature range of different
measurements. Even so, an equal or greater contribu-
tion to the variability was attributed to poor methodol-
ogy, including poor data collection and improper kinetic
analysis. Kerogen and coal kinetics are commonly and
perhaps usually determined using ramped-temperature
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM

experiments. Kissinger’s method (Kissinger 1957) is a
very easy and common method of extracting A and E
values from multiple heating rates. Braun et  al. (1991)
and Burnham (2000) extended this method to obtain esti-
mates of the Gaussian E distribution parameter, reaction
order and nucleation-growth parameters for distributed
and sigmoidal reactivities from the reaction profile widths
and asymmetries.
Reactivity diversity can be described by a distribution
in A, E or both (Burnham and Braun 1999). A pseudo-nth-
order reaction is equivalent to a Gamma distribution in A.
A variety of continuous distributions have been used for
E, with a Gaussian distribution being the most common.
However, the reactivity distribution is rarely symmetric,
so a Gaussian distribution of pseudo-nth-order reactions
is superior. Alternatively, a Weibull distribution is very
flexible. However, multiple reaction groups are usually
present, which require parallel continuous distribution
models. Using a discrete E distribution is more powerful
and is the prevalent choice for petroleum systems mod-
eling. Isoconversional methods are equally powerful and
suitable for complex kinetic analysis of fossil fuels (Bai
et  al. 2015b). Tiwari and Deo (2012b) concluded that the
advanced isoconversional method yielded activation
energies as a function of conversion in the range of 93–
245 kJ mol − 1. The decomposition process can be viewed as
consisting of multiple parallel reactions with individual
activation energies. Maximum uncertainties in activa-
tion energies computed using the advanced isoconversion
method were about 10% of the energy values calculated.
There is a wide variation in the literature for the values of
activation energy. The type of oil shale and type of pyroly-
sis, whether isothermal or non-isothermal, are influential
factors for the discrepancies.
Because the pyrolysis of type I kerogen is usually
dominated by a single apparent activation energy, its
pyrolytic character is more consistent as a function of
the extent of conversion. The Rock-Eval temperature
of maximum reaction rate (Tmax) is roughly constant at
440–445°C as a function of geological maturity until all
oil potential is exhausted. Unfortunately, the Rock-Eval
Tmax scale was established prior to accurate temperature
measurements, and those values are lower than the true
Tmax by 35–38°C. Nevertheless, if one simply makes that
correction and assumes A = 1014 s − 1, one obtains a mean
activation energy of about 226.08 kJ mol − 1. This is within
the range of plausible values. Type II kerogens typi-
cally have distributed reactivity roughly consistent with
a Gaussian distribution of 2% in the activation energy.
The Rock-Eval Tmax of immature type II kerogens is typi-
cally approximately 420–425°C. Again assuming A = 1014
M.A. Raja et al.: Oil shale pyrolysis and kinetics      13
s − 1, one obtains a mean activation energy of a little less
than 221.90  kJ mol − 1. In fact, it is equally plausible that
the initial mean activation energy for both type I and II
kerogens is 221.90 kJ mol − 1 and that the difference in the
reactivity is due to A = 5 × 1013 s − 1 for type I kerogen and
A = 1.3 × 1014 s − 1 for type II kerogen, but whether the reac-
tivity difference is absorbed into A or E makes no practi-
cal difference for modeling oil shale retorting. Because of
the reactivity distribution for type II kerogens, the Rock-
Eval Tmax increases to 440–445°C by the end of oil genera-
tion. After the completion of oil generation, the Rock-Eval
Tmax for both type I and II kerogens continue to increase,
reflecting the increasingly difficult demethylation reac-
tions as semi-coke moves toward graphite.
Among the most important factors are the type of
sweeping gas, heating rate employed and the final pyrol-
ysis temperature; the oil yield and the extent of any oil
cracking are greatly influenced by the peak pyrolysis
temperature. The main objective of the sweeping gas is
the removal of generated products (Omar et  al. 2010).
Omar (2006) described that the reaction environment
during pyrolysis of oil shale has a significant effect on
the liquid product distribution and its sulfur content. The
total sample weight loss and the oil yield are affected by
the sweeping medium. The type of sweep gas or sweep-
ing mixture employed during pyrolysis influenced the
percentages of fractions and the sulfur content. Ekinci
et  al. (1996) investigated the influences of steam and
nitrogen sweep gases in a fixed bed reactor on the oil
yield and product distribution. They reported an increase
in oil yield and higher concentration of n-alkane and an
increase in the aromatic character of the non-paraffinic
product materials. Hershkowitz et  al. (1983) reported a
20% increase in kerogen conversion when hydrogen is
used as pyrolysis atmosphere in their study of Colorado
oil shale; in addition, they reported 90% organic carbon
conversion in the 549.85–599.85°C temperature range.
Eastman and Schlinger (1964) obtained 126% (vol.%) oil
yield of Fischer assay when they used high hydrogen pres-
sure during pyrolysis. El-hafri et  al. (2000) studied the
effect of steam on product distribution of oil shale pyroly-
sis. They found that oil yield increases with the addition of
steam to nitrogen sweeping gas.
4 G
 raphical methods to determine
oil shale pyrolysis kinetics
An understanding of kinetic parameters, such as activa-
tion energy and the frequency factor, may be obtained
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM
14      M.A. Raja et al.: Oil shale pyrolysis and kinetics
by various approaches (Wanjun et  al. 2003). Graphi-
cal methods are often used to evaluate goodness-of-fit
(Xiaoshu et  al. 2014). Table 2 summarizes the work of
investigators who used graphical methods to analyze
kinetics of oil shale. Rajeshwar (1981) analyzed the weight
loss data by direct Arrhenius, Coats Redfern and Freeman
Carroll techniques. The direct Arrhenius method and the
Coats Redfern method are found to be convenient for the
analysis of the data. The Freeman Carroll method is rather
more tedious and moreover is of little use in analyzing
the early stages of the decomposition. The analysis of the
kinetic data by the Freeman Carroll method is shown to
confirm the validity of assuming first-order kinetics for
the thermal decomposition of Green River oil shale. Doǧan
and Uysal (1996) used the integral method, and non-iso-
thermal TGA data, to determine the overall kinetics of
pyrolysis of three Turkish oil shales. Essentially, either
differential or integral methods are applied to analyze
non-isothermal experimental data. The integral method is
generally accepted as more accurate, but even the approx-
imations used in integration techniques are effective in
the final results for kinetic parameters. Khraisha (1998)
described the kinetic parameters obtained by weight
vs. time data at temperatures 374.85–474.85°C using the
Arrhenius method for two different Jordan shales. The rate
of kerogen decomposition can adequately be described by
a simple mechanism involving first-order kinetics. Possi-
bly, two consecutive first-order reactions may take place
during the decomposition process.
Shuyuan and Changtao (2004) carried out pyro­
lysis experiments on oil shale samples from Fushun and
Maoming using a TGA analyzer. The kinetic parameters
of oil shale pyrolysis were determined from TGA data by
using the integral method, differential method, Friedman
procedure, maximum rate method and parallel first-order
reaction model. A special regression function was used
to calculate the pyrolysis reaction rate. Relative regres-
sion coefficients by the integral method and differential
method were larger than 0.99. Shabbar et al. (2011) inves-
tigated the integral method, direct Arrhenius plot method
and method of approximate temperature integral. The cal-
culated kinetic parameter results were compared with the
experimental curve of TGA analysis and also among them-
selves. A root mean square (RMS) error was calculated for
each method and the integral method model was found
to be the closest to experimental data. Torrente and Galán
(2001) studied the kinetics of the thermal decomposition
of Spanish oil shale using isothermal and non-isothermal
TGA. Three methods were applied for the determination of
the kinetic parameters, the direct Arrhenius plot method,
the integral method and the differential method. Among
the three models used in non-isothermal conditions, the
integral method yields lower deviation and hence pro-
vides a better fit of the data.
5 C
 hallenges in determination of oil
shale pyrolysis kinetics
In the depletion of oil and natural gas resources, chal-
lenges related to the conversion of solid fossil fuels (coal,
shale and peat) into motor fuels and chemicals have
become of considerable practical importance. In particu-
lar, interest has quickened in the processing of oil shale,
whose geological reserves are in abundance. Some of the
experimental and modeling challenges of oil shale pyro­
lysis kinetics are described below.
5.1 Experimental challenges
Isothermal conditions are often difficult to ensure during
the entire course of reactions that occur at high temper-
ature. This difficulty is the result of the initial heat-up
period required to reach the isothermal reaction tem-
perature (Campbell et al. 1978). A part of the sample may
change while the sample is heated to the desired tempera-
ture, especially at the degradation of high polymers; this
initial structure change in the sample complicates the
isothermal data and makes it difficult to analyze (Ozawa
1965). The time required for the test samples to attain the
reaction temperature introduces a significant error in the
kinetic data, especially at elevated temperatures when the
heat-up time becomes a significant fraction of the total
reaction period (Rajeshwar 1981, Doǧan and Uysal 1996,
Jaber et al. 1999).
Thermal methods providing information about net
results of mass loss and calculation of kinetic parameters
are based on simplifying assumptions, which do not cor-
respond to the complex chemical reactions in the thermal
degradation of the oil shale (Kök and Pamir 2000). The
use of TGA to determine kinetic parameters for the pyroly-
sis of oil is complicated in that the oil shales are a complex
mixture of kerogen and a wide range of minerals. In addi-
tion, the decomposition of the oil shales represents a
large number of reactions in parallel and series whilst the
TGA measures the overall weight loss due to these reac-
tions. Therefore, the TGA provides general information
on the overall reaction kinetics rather than individual
reactions, and therefore the activation energies derived
from TGA should be termed apparent activation energies
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM

(Ahmad and Williams 1998). Accurate isothermal and
non-­isothermal operations at large scale are not possible.
5.2 C
 hallenges in determination of oil shale
pyrolysis mechanism and kinetics
Oil shale pyrolysis is a very complicated thermochemical
conversion process involving extremely complex reactions
and the end products depend on numerous factors (Tiwari
and Deo 2012b, Xiaoshu et  al. 2014). Since kerogen is a
complex heterogeneous mixture of organic compounds
with an approximate formula of C200H300SN5O11 (Khraisha
1998), its decomposition includes a very complex reac-
tion, or group of reactions, that results in the production
of different types of products ranging from a single carbon
atom per molecule (CH4) to compounds containing more
than 30 atoms per molecule (Khraisha 1998). The wide
variety of products obtained by the thermal processing of
oil shale pose difficulties in the determination of the kinet-
ics and mechanism of pyrolysis (Kök and Pamir 2000).
Apart from the decomposition of the oil shale
kerogen, there are a variety of other reactions that are
brought about by application of heat. For example, the
presence of host minerals in the relevance of measured
parameters to on-field applications of oil shale matrix
significantly complicates the thermal behavior. These
reactions are critically dependent on the nature of the
atmosphere surrounding the shale. The complexity of the
kerogen decomposition process is compounded by the
presence of secondary reactions which include secondary
cracking and polymerization. The rate of heating and the
possible catalytic effects of mineral matter are additional
complicating factors (Rajeshwar et al. 1979). The complex
reaction schemes of oil shale pyrolysis have been pre-
sented in Table 1.
Different mechanisms, with various degrees of com-
plexity, have been proposed to describe oil shale pyrolytic
behavior. Generally, these are based on tracking interme-
diate, primary and secondary pyrolysis products, often
without a clear identification of all of the products in any
of the classes being tracked. Kinetic parameters have been
calculated by fitting experimental data to an assumed
semi-empirical model of the pyrolysis process. There is,
thus, not surprisingly a large variation in reported kinetic
parameters (activation energies and frequency factors).
Consequently, the results are often of limited general-
ity, applicable to the particular experimental conditions
studied, often potentially influenced by heat and mass
transfer limitations on observed kinetics (Vahur and Eric
2013).
M.A. Raja et al.: Oil shale pyrolysis and kinetics      15
The structure and composition of the organic matter
(kerogen) are not yet well established. The product dis-
tribution shows a uniform trend that is not quantitative.
Secondary reactions occur during the process through
complicated mechanisms.
5.3 Challenges of graphical methods
Graphical methods are although simple, an obvious dis-
advantage with these methods is to describe observations
of the data only without showing the underlying mecha-
nism by which the data are produced. Another relevant
disadvantage, which is directly tied to these methods, is
that graphical methods generally suffer from poor statis-
tical properties (Burnham and Braun 1999, Xiaoshu et al.
2014).
5.4 Modeling challenges
The structure and composition of the organic matter
(kerogen) are not yet well established. Thus the mod-
eling of oil shale pyrolysis kinetics is extremely compli-
cated (Bai et  al. 2014). Vandenbroucke (2003) described
that the name kerogen, in opposition with usual chemis-
try nomenclature, does not represent a substance with a
given chemical composition. Indeed kerogen is a generic
name, in the same sense as lipids or proteins. Vandenb-
roucke and Largeau (2007) explained that the general term
“kerogen” cannot represent a defined chemical structure.
In fact, “true” macromolecular components of kerogen
cannot be represented because their molecular composi-
tion is not known. Only chemical structures like those of
synthetic polymers, where a small sub-unit repeats the
whole structure, can be represented by a detailed chemi-
cal structure. Chemical structural information about
kerogen is critical for assessing the thermal maturity of
source rocks and evaluating the HC generation potential.
However, knowledge of the detailed chemical changes of
kerogen with increasing thermal maturity remains insuffi-
cient (Vandenbroucke 2003, Vandenbroucke and Largeau
2007).
A single approach to the thermal processing of oil
shale is difficult to determine, and process parameters
and implementation are difficult to optimize because of
differences in the composition and properties of oil shale
from various deposits. The composition and quantitative
ratios between the pyrolysis products of oil shale depend
on many factors: the type and origin of the organic matter
of oil shale, the final process temperature, the rate of
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM
16      M.A. Raja et al.: Oil shale pyrolysis and kinetics
heating, the residence time in a high temperature zone,
grain size, composition and pressure (Strizhakova and
Usova 2008).
Many mechanisms could be postulated. Parameters in
mathematical models can force a good fit, but this does
not mean that the model is unique. Reliable kinetic data
are essential for the accurate mathematical modeling of
various on-surface and in situ oil shale processes. Due to
the very complex nature of the organic matter in oil shale,
unraveling the kinetics has not been straightforward. An
accurate kinetic model that is able to adequately represent
the mechanism of kerogen conversion to generated prod-
ucts is necessary.
Most of the developed kinetic models by the research-
ers have limited application.
6 Discussion and summary
Since an accurate Arrhenius model equation is inherently
nonlinear in nature allowing no exact analytical solu-
tion, numerical techniques become necessary to find the
solutions and the parameters (e.g. Youtsos et  al. 2013).
To determine an appropriate valid reaction model, one
of several available methods, which could be sigmoidal,
distributed reactivity or multiple reactions depending
on the material are used. Valuable insights are avail-
able through isoconversional (Burnham and Braun 1999,
Brown et al. 2000, Vyazovkin et al. 2011), extended Kiss-
inger (Burnham and Braun 1999, Burnham 2000) and gen-
eralized reaction profile (Gotor et al. 2000) methods. Also,
look for inflection points diagnostic of multiple reactions.
One cannot determine the correct model from sparse data
where the maximum conversion varies substantially with
temperature.
Shale oil produced by pyrolysis is the result of several
physical and chemical reactions occurring in series and
parallel (Jaber and Probert 1999, 2000, Jaber et al. 1999).
The simple model is usually inappropriate for the simula-
tion of pyrolysis over a full range of conditions, so more
detailed representation of the pyrolysis process is neces-
sary by combining parallel and consecutive reaction steps
(Bar et al. 1988, Schenk and Horsfield 1993, Shuyuan and
Changtao 2003, 2004, Tiwari and Deo 2012b). Kinetic
parameters calculated from the parallel model will
provide more reasonable information for the oil shale
retorting process and for recognizing the chemical struc-
ture of oil shale kerogen (Bar et al. 1988, Skala et al. 1989,
Schenk and Horsfield 1993, Shuyuan and Changtao 2003,
2004, Tiwari and Deo 2012b, Bai et al. 2015a). Skala et al.
(1989) investigated three alternative mathematical models
based on single-step, two-step and multistep (MSM) reac-
tion schemes. The advantages of the MSM in predicting
the actual heat consumption rates, and in predicting the
pyrolysis effects in both non-isothermal and isothermal
regimes, appear to be of significance. Shuyuan and Chang-
tao (2003) developed a kinetic model which assumes
several parallel first-order reactions with changed acti-
vation energies and frequency factors to describe the oil
shale pyrolysis.
The correct reaction model was rarely close to first
order, with most kerogens requiring a reactivity distribu-
tion (Schaefer et al. 1990, Braun et al. 1991, Tegelaar and
Noble 1994, Peters et  al. 2006) as for coals, and some
well-preserved algal kerogens requiring a sigmoidal reac-
tion model (Burnham et al. 1996, Peters et al. 2015). The
decomposition temperatures of immature to early-mature
coals and kerogens vary by only about 50°C, which corre-
sponds to a change of about 10% in the activation energy
or a factor of about 16 in the frequency factor. The range
of variability in A and E is somewhat larger due to real
mechanistic compensation effects. The range of frequency
factors is most likely in the 1012–1016 s − 1 range, with most
values within the middle half of that range. The range of
variability in the principal activation energy is from about
200 to 242 kJ mol − 1, with most samples being in the middle
half of that range (Burnham et al. 1987, 1988, Sundarara-
man et al. 1988, Pepper and Corvi 1995). There is little evi-
dence that the activation energy for conversion of kerogen
to bitumen, oil and gas is significantly different from the
activation energy for conversion to volatile HCs ( < C32 or
so). Realistic kinetic parameters for both kerogen conver-
sion and oil creation are constrained by geological obser-
vation of their respective stabilities (Pepper and Corvi
1995).
Only a few attempts have been made to develop an
advanced model of oil shale pyrolysis, relevant to the
design of industrial scale oil shale retorting (Wallman
et al. 1981, Burnham and Braun 1984). Modeling the pyrol-
ysis kinetics of oil shale for understanding the decompo-
sition mechanisms and kinetic parameters is significant
for designing and operating an energy efficient pyrolysis
process (Skala et al. 1989, Tiwari and Deo 2012b, Xiaoshu
et  al. 2014). The developed kinetic models have been
complex in nature, utilizing intermediates formed from
kerogen initial decomposition and the secondary reac-
tions to produce gases and shale oil products (Hubbard
and Robinson 1950, Khraisha 1998, Xiaoshu et al. 2014).
Lewis and Braun (1981) described a mathematical model,
developed at the Lawrence Livermore National Laboratory
(LLNL), which is based on an improved understanding of
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM
 M.A. Raja et al.: Oil shale pyrolysis and kinetics      17

the physical and chemical processes involved in retort- For a constant heating rate β = dT/dt Eq. (6) may be
ing. It has since been modified to calculate the retorting written as
of shale moving through a retort, as required for a con- −E
dα
tinuous process in surface equipment. First, the model β = Ae RT ( t ) f ( α). (7)
dT
was modified to allow the reacting shale particles to move 
through the retort at a specified linear velocity, as required
The integral form of this equation after separating
for a continuous process in surface equipment. Second,
variables is
intraparticle thermal conduction was incorporated into
the model. Then the shale temperature and composition α*
dα 1
T* −E
1
were determined as a function of radial position within a ∫ Af ( α) = β ∫ e RT ( t )
dT =  I ( E ,T ) .
β α α 
(8)
particle, both by the gas-solid heat transfer coefficient and 0 0

by the shale thermal conductivity. The developed math-
The assumptions that the reaction model does not
ematical model was considered as a powerful tool in the
depend on heating rates and is constant for a small con-
development of retorts, especially in understanding the
version interval lead to the integral form of the rate law:
combined effects of heat-transfer processes and the ther-
modynamics and kinetics of the many chemical reactions. αend − αstart I ( E α ,Tα )
= . (9)
Thus, it was believed that application of this mathemati- Af ( α) β

cal model may contribute to improve surface processing.
Tiwari and Deo (2012b) showed that the advanced iso- These assumptions suggest that the integral at any
conversion method provides a combined pre-exponential particular conversion should be the same for all heating
factor and reaction model as a function of conversion. The programs and be a function of the time-temperature rela-
RMS errors between the experimental and model values tionship. According to this, for a set of N experiments
for the different approaches were compared. The isocon- carried out at different heating programs, the activation
version approach produced the lowest RMS values in both energy is determined at any particular level of conversion
rates and cumulative conversion. by minimizing the following function:
Globally, single-stage decomposition was assumed 2
in deriving kinetic rate expressions (Tiwari and Deo 1 N  I ( E ,T ) 
2012b).
χ (E) =
2
∑∑ 1 −
i
 ,
N ( N − 1) i=1 j ≠i  I j ( E ,T ) 
(10)

→ Decomposition products. (3)
Kerogen 
where the subscripts i and j represent two experiments
The conversion of solid matter in shale (kerogen) to performed under different heating programs. The trap-
products from TGA weight loss data is defined as (Tiwari ezoidal rule is used to evaluate the integral numerically
and Deo 2012b) and the minimization procedure is repeated for each value
of α to find the dependence of activation energy on the
W0 − Wt
α= , (4) extent of conversion. The activation energy distribution
W0 − W∞
 obtained in Eq. (10) may be used to determine [A·f(α)] as
a function of α.
where W0 is the initial weight of the sample (mg), Wt is The experimental rate and conversion data can be
the weight of the sample at time t (mg), W∞ is the weight reconstructed based on the model parameters using
of the sample at the end of the experiment (mg) and α is Eq. (7):
conversion.
 E 
In general, the rate of decomposition can be expressed ln  βi(d α dT )αi  = ln  Af ( α) −  α  . (11)
using the non-parametric kinetic equation  RTα,i 

dα
= f (T ) ⋅ f ( α). (5) Kinetic models can be used to extrapolate to non-
dt
 experimental rates. Slow pyrolysis that is likely during in
situ oil shale production and high rates of flash pyroly-
Using the Arrhenius expression leads to the following:
sis are of interest. The assumptions of the isoconversion
−E method [Eq. (9)] allow calculating the temperature to
dα
= f (T ) f ( α) = Ae RT ( t ) f ( α). (6) reach a level of conversion at extrapolated heating rates
dt 

Brought to you by | University of Colorado - Boulder

Authenticated
Download Date | 4/1/17 10:02 AM
 18      M.A. Raja et al.: Oil shale pyrolysis and kinetics

using the following mathematical equivalency (Vyazovkin (Burnham and Braun 1999, Brown et  al. 2000, Burnham
and Lesnikovich 1992, Tiwari and Deo 2012b): 2000, Gotor et al. 2000, Vyazovkin et al. 2011, 2014, Peters
et al. 2015).
I ( E α ,Tα,i ) I ( E α ,Tα, j )
= . (12)
βi βj

Eq. (12) was used to estimate the temperature at which 7 Conclusions and recommendations
the material starts to convert. The procedure for recon-
struction was then used to obtain conversions and rates at
for future work
extrapolated conditions (Tiwari and Deo 2012b).
The study carried out in this research work concludes the
Tiwari and Deo (2012b) described two distinct data
following points.
analysis approaches, advanced isoconversional method
Py-GC-MS is a favorable method to chemically char-
and parameter fitting, using the TGA dataset of the
acterize the main structural skeleton in the kerogen at
Green River oil shale. RMS errors between the model
the molecular level. Any improvement in knowledge of
and experimental data were the lowest for the isoconver-
kerogen composition and structure may increase the reli-
sional method, but the distributed reactivity models also
ability of kinetic models in predicting the amount and
produced reasonable results. Simulated decomposition
composition of generated petroleum products.
rates and onset temperatures shift to higher temperatures
Classic graphical methods cannot unambiguously
at higher heating rates. When using parameter fitting
measure and estimate kinetic parameters due to the math-
approaches, a number of models produce similar results
ematical complexity. The simple model is usually inap-
making model choice difficult. A MATLAB-based compu-
propriate for the simulation of pyrolysis over a full range
tational method described in the text was used. Results
of conditions. Kinetic parameters calculated from the
of the model comparisons with the experimental data are
parallel model will provide more reasonable information
shown in Figure 4. The agreement between the model and
for the oil shale retorting process and for recognizing the
the experimental data is good over most of the conversion
chemical structure of oil shale kerogen.
range, and for all the rates. The extrapolated results are
Advanced isoconversion methods would be appro-
not all consistent with some experimental results.
priate for the calculation of the distribution of activation
To derive valid kinetic parameters, one must follow
energies for multiple reactions involved in the decomposi-
the following steps: decouple temperature and conver-
tion of complex material such as kerogen to products.
sion, eliminate or measure thermal gradients, determine
The range of variability in the principal activation
the reaction model, fit the data, and confirm agreement,
energy is from about 200 to 242 kJ mol − 1, with most samples
which are summarized from guidance in the literature
being in the middle half of that range, while the range of
frequency factors is most likely in the 1012–1016 s − 1 range,
1.0 with most values within the middle half of that range.
Some recommendations that are suggested for further
Expt- 0.5 C/min research work to resolve differences in reported kinetics
0.8 Simul-0.5 C/min
Expt-1 C/min are as follows.
Simul-1 C/min
Expt-2 C/min An important shortfall of the presented kinetic models
Normalized conversion

Simul-2 C/min
0.6 Expt-5 C/min
Simul-4 C/min
is their limited application. There is a need of such models
Expt-10 C/min
Simul-10 C/min
that are able to fit the data and extrapolate beyond the
Expt-20 C/min
Simul-20 C/min range of the data, generic and application independent.
0.4
Isoconversion models are in theory “kinetic model”
Expt-50 C/min
Simul-50 C/min

free and their applicability to the decomposition of a

0.2 complex material such as kerogen is excellent. However,
their applicability in reproducing multistep kinetics has
been questioned. More attempts are required to develop
0.0
200 250 300 350 400 450 500 550 an advanced model of oil shale pyrolysis, relevant to the
Temperature (°C) design of industrial scale oil shale retorting.
Figure 4: Experimental and simulated conversion profiles at dif-
The study of mineralogy suggests that additional
ferent heating rates using the advanced isoconversional method reactions, dehydration and dehydroxylation, need to
(Tiwari and Deo 2012b). be included in the complete pyrolysis mechanism. The

Brought to you by | University of Colorado - Boulder

Authenticated
Download Date | 4/1/17 10:02 AM

mathematical models need to be improved to account for
the thermodynamic behavior of the products formed.
Nomenclature
W0 Initial weight of the sample (mg)
Wt Weight of the sample at time t (mg)
W∞ Weight of the sample at the end of the experiment (mg)
α Conversion
β Constant heating rate (°C min − 1)
K* Kerogen and bitumen originally present in oil shale
P All volatilized products formed
B All non-volatilized products
f1 and f2 Stoichiometric coefficients
t Time
N Number of heating rates
I Integral symbol
Tα Temperature at conversion α
R Gas constant (8.314 kJ mol − 1 K − 1)
T Temperature (K)
Tmax Temperature of maximum reaction rate
E Apparent activation energy (kJ mol − 1)
Eα Activation energy at conversion α
f(α) Reaction model
f(T) Temperature dependence of the reaction rate
Acknowledgments: This work was financially sup-
ported by the National Basic Research Program of China
(2014CB744306), the National Natural Science Fund for
Distinguished Young Scholars (21425625), and supported
by the External Cooperation Program of B1C, Chinese Acad-
emy of Sciences (122111KYS820150017) and the National
Project of International Cooperation (2013DFG62640).
References
Ahmad N, Williams PT. Influence of particle grain size on the yield
and composition of products from the pyrolysis of oil shales.
J Anal Appl Pyrol 1998; 46: 31–49.
Bai F, Sun Y, Liu Y, Liu B, Guo M, Lü X, Guo W, Li Q, Hou C, Wang Q.
Kinetic investigation on partially oxidized Huadian oil shale by
thermogravimetric analysis. Oil Shale 2014; 31: 377.
Bai F, Guo W, Lü X, Liu Y, Guo M, Li Q, Sun Y. Kinetic study on the
pyrolysis behavior of Huadian oil shale via non-isothermal
thermogravimetric data. Fuel 2015a; 146: 111–118.
Bai F, Sun Y, Liu Y, Li Q, Guo M. Thermal and kinetic characteristics
of pyrolysis and combustion of three oil shales. Energy Convers
Manage 2015b; 97: 374–381.
Bar H, Ikan R, Aizenshtat Z. Fossil fuels and synthetic polymers:
isothermal pyrolysis kinetics as indication of structural
resemblance. J Anal Appl Pyrol 1988; 14: 73–79.
Behar F, Lorant F, Lewan M. Role of NSO compounds during primary
cracking of a Type II kerogen and Type III lignite. Org Geochem
2008; 39: 1–22.
M.A. Raja et al.: Oil shale pyrolysis and kinetics      19
Berna H, Tayfun B, Serhat A. Experimental and numerical simulation
of oil recovery from oil shales by electrical heating. Energ Fuel
2008; 22: 3976–3985.
Braun RL, Christiansen DE, Diaz JC, Lewis AE. Results of
mathematical modeling of oil shale retorting in an
aboveground, internal combustion retort. Fuel Process Technol
1984a; 9: 125–138.
Braun RL, Diaz JC, Lewis AE. Results of mathematical modeling of
modified in-situ oil shale retorting. Soc Pet Eng J 1984b; 24:
75–86.
Braun RL, Burnham RL, Reynolds AK, Clarkson JE. Pyrolysis kinetics
for lacustrine and marine source rocks by programmed
micropyrolysis. Energ Fuel 1991; 5: 192–204.
Brendow K. Global oil shale issues and perspectives. Oil Shale
2003; 20: 81–92.
Brown ME, Maciejewski M, Vyazovkin S, Nomen R, Sempere J,
Burnham A, Opfermann J, Strey R, Anderson HL, Kemmler A,
Keuleers R, Janssens J, Desseyn HO, Li CR, Tang TB, Roduit
B, Malek J, Mitsuhashi T. Computational aspects of kinetic
analysis: part A. The ICTAC kinetics project-data, methods and
results. Thermochim Acta 2000; 355: 125–143.
Burnham AK. Pyrolysis kinetics and composition for Posidonia
shale. LLNL Report UCRL-ID-105, 1990.
Burnham AK. Application of the Šesták-Berggren equation to organic
and inorganic materials of practical interest. J Therm Anal Cal
2000; 60: 895–908.
Burnham AK. Chemistry and kinetics of oil shale retorting. In:
Ogunsola OI, Hartstein AM, Ogunsola O, editors, Oil shale: a
solution to the liquid fuel dilemma. ACS Symposium Series
1032, Washington, DC: American Chemical Society, 2010.
Burnham AK. A simple kinetic model of oil generation, vaporization,
coking, and cracking. Energ Fuel 2015; 29: 7156–7167.
Burnham AK, Braun RL. General kinetic model of oil shale pyrolysis.
Lawrence Livermore National Laboratory, 1984.
Burnham AK, Braun RL. Global kinetic analysis of complex materials.
Energ Fuel 1999; 13: 1–22.
Burnham AK, Braun RL, Greaa HR, Samoun AM. Comparison of
methods for measuring kerogen pyrolysis rates and fitting
kinetic parameters. Energ Fuel 1987; 1: 452–458.
Burnham AK, Braun RL, Samoun AM. Further comparison of methods
for measuring kerogen pyrolysis rates and fitting kinetic
parameters. In: Nattavelli L, Novelli L, editors, Advances in
organic geochemistry 1987, Part II. Oxford: Pergamon Press,
1988: 839–846.
Burnham AK, Braun RL, Sweeney JJ, Reynolds JG, Vallejos C,
Talukdar S. Kinetic modeling of petroleum formation in the
Maracaibo basin: Final Report. DOE Report DE92001051,
1992a.
Burnham AK, Samoun AM, Reynolds JG. Characterization of
petroleum source rocks by pyrolysis-mass spectrometry gas
evolution profiles. Lawrence Livermore National Laboratory
Report UCRLID-111012, 1992b.
Burnham AK, Braun RL, Coburn TT, Sandvik EI, Curry DJ, Schmidt BJ,
Noble RA. An appropriate kinetic model for well-preserved algal
kerogens. Energ Fuel 1996; 10: 49–59.
Bushnev DA, Burdelnaya NS. Kerogen: chemical structure and
formation conditions. Russ Geol Geophys 2009; 50:
638–643.
Campbell JH, Koskinas GH, Stout ND. Kinetics of oil generation from
Colorado oil shale. Fuel 1978; 57: 372–376.
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM
20      M.A. Raja et al.: Oil shale pyrolysis and kinetics
Castelli A, Chiaramonte MA, Beltrame L, Carniti, Del BA, Stoppe F.
Thermal degradation of kerogen by hydrous pyrolysis. A kinetic
study. Org Geochem 1990; 16: 75–82.
Coats AW, Redfern J. Kinetic parameters from thermogravimetric
data. Nature 1964; 201: 68–69.
Coburn TT, Taylor RW, Morris CJ, Duval V. Isothermal pyrolysis
and char combustion of oil shales. In: Proc. 21st Oil Shale
Symposium, Beijing, China, 1988: 245–252.
Deepak ST, Eric HN. Kinetics of pyrolsis of Moroccan oil shale by
thermogravimetry. Ind Eng Chem Res 1987; 26: 1351–1356.
Doǧan ÖM, Uysal BZ. Non-isothermal pyrolysis kinetics of three
Turkish oil shales. Fuel 1996; 75: 1424–1428.
Dyni JR. Geology and resources of some world oil-shale deposits. Oil
Shale 2003; 20: 193–252.
Eastman B, Schlinger WG. US Patent 3117072, 1964.
Ekinci E, Citiroglu M, Akar A, Putun E, Snape CE. Hydropyrolysis and
catalytic hydropyrolysis of oil shales. Hydrogen Energy System
1996; 295: 281–286.
El-hafri, Mokhlisse KA, Ben CM. Yields and composition of oil
obtained by isothermal pyrolysis of the Moroccan (Tarfaya)
oil shales with steam or nitrogen as carrier gas. J Anal Appl
Pyrolysis 2000; 56: 207–218.
Farouq MS, Ahmed UO, Talal AW, Yahya AW, Insa AM. Deep oxidative
desulfurization of liquid fuels. Rev Chem Eng 2014; 30:
337–378.
Freund H, Clouse JA, Otten GA. Effect of pressure on the kinetics of
kerogen pyrolysis. Energ Fuel 1993; 7: 1088–1094.
Freund H, Walters CC, Kelemen SR, Siskin M, Gorbaty ML, Curry DJ,
Bence AE. Predicting oil and gas compositional yields via chemical
structure-chemical yield modeling (CS-CYM): Part I – concepts and
implementation. Org Geochem 2007; 38: 288–305.
Gersten J, Fainberg V, Hetsroni G, Shindler Y. Kinetic study of the
thermal decomposition of polypropylene, oil shale, and their
mixture. Fuel 2000; 79: 1679–1686.
Gotor FJ, Criado JM, Malak J, Koga N. Kinetic analysis of solid-state
reactions: the universality of master plots for analyzing
isothermal and nonisothermal experiments. J Phys Chem A
2000; 104: 10777–10782.
Haifeng J, Lihua S, Zhiqiang C, Jie C, Li Z, Mingyue Z, Meijuan H, Jianing
L, Junfeng L. Influence of pyrolysis condition and transition metal
salt on the product yield and characterization via Huadian oil
shale pyrolysis. J Anal Appl Pyrol 2015; 112: 230–236.
Hershkowitz F, Olmstead WM, Rhodes RP, Rose KD. Molecular
mechanism of oil shale pyrolysis in nitrogen and hydrogen
atmospheres. Am Chem Soc 1983; 230: 301–316.
Hubbard AB, Robinson WE. A thermal decomposition study of Colorado
oil shale. Report of investigation. US Bureau of Mine, 1950.
Jaber JO, Probert SD. Pyrolysis and gasification kinetics of Jordanian
oil-shales. Appl Energ 1999; 63: 269–286.
Jaber JO, Probert SD. Non-isothermal thermogravimetry and
decomposition kinetics of two Jordanian oil shales under different
processing conditions. Fuel Process Technol 2000; 63: 57–70.
Jaber JO, Probert SD, Williams PT. Evaluation of oil yield from
Jordanian oil shales. Energy 1999; 24: 761–781.
Jian VL, Steven WJ, Yanfa Y, Dean HL. Measuring temperature-
dependent activation energy in thermally activated processes:
a 2D Arrhenius plot method. Rev Sci Instrum 2010; 81, pp.
033910(1)–033910(5). doi: 10.1063/1.3361130.
Jiang XM, Han XX, Cui ZG. New technology for the comprehensive
utilization of Chinese oil shale resources. Energy 2007; 32:
772–777.
Jinhu D, He L, Desheng M, Jinhua F, Yuhua W, Tiyao Z. Discussion
on effective development techniques for continental tight oil
in China. Petroleum Exploration and Development 2014; 41:
217–224.
Johannes I, Zaidentsal A. Kinetics of low-temperature retorting of
kukersite oil shale. Oil Shale 2008; 25: 412–425.
Johnson WF, Walton DK, Keller HH, Crouch EJ. In situ retorting
of oil shale rubble: a model of heat transfer and product
formation in oil shale particles. Colo Sch Mines Quart 1975;
70: 237–272.
Karabakan A, Yürüm Y. Effect of the mineral matrix in the reactions
of oil shales: 1. Pyrolysis reactions of Turkish Göynük and US
Green River oil shales. Fuel 1998; 77: 1303–1309.
Katti DR, Upadhyay HB, Katti KS. Molecular interactions of kerogen
moieties with Na-montmorillonite: an experimental and
modeling study. Fuel 2014; 130: 34–45.
Khraisha YH. Kinetics of isothermal pyrolysis of Jordan oil shales.
Energy Convers Manage 1998; 39: 157–165.
Maqsood K, Aditi M, Abulhassan A, Kajari K and Saibal G. Cryogenic
carbon dioxide separation from natural gas: a review based
on conventional and novel emerging technologies. Reviews in
Chemical Engineering 2014; 30: 453–477.
Kissinger HE. Reaction kinetics in differential thermal analysis. Anal
Chem 1957; 29: 1702–1706.
Kök MV, Pamir MR. Comparative pyrolysis and combustion kinetics
of oil shales. J Anal Appl Pyrol 2000; 55: 185–194.
Le DTV, Bostrom NW, Burnham AK, Kleinberg RL, Pomerantz AE, Allix
P. Green river oil shale pyrolysis: semi-open conditions. Energ
Fuel 2013; 27: 6447–6459.
Lewis AE, Braun RL. Retorting and combustion processes in surface
oil-shale retorts. J Energ 1981; 5: 355–361.
Martins MF, Salvador S, Thovert JF, Debenest G. Co-current
combustion of oil shale – Part 1: characterization of the solid
and gaseous products. Fuel 2010; 89: 144–151.
McKee RH, Lyder EE. The thermal decomposition of shales. I – heat
effects. Ind Eng Chem 1921; 13: 613–618.
Mohammad AH, Omar AA, John R, Sam K, Adnan AH, Khalid T,
Abdurrahman S, Richelieu B. Effect of demineralization and
heating rate on the pyrolysis kinetics of Jordanian oil shales.
Fuel Process Technol 2011; 92: 1805–1811.
Omar SAA. Sweeping gases influence on liquid product and sulfur of
oil shale. International Conference on Oils Shale: “Recent Trends
in Oil Shale”, 7–9 November 2006, Amman, Jordan. 2006. Paper
No. rtos-A115.
Omar SAA, Matouq M, Anbar Z, Khaleel AM, Abu-Nameh E. Oil shale
pyrolysis kinetics and variable activation energy principle. Appl
Energ 2010; 87: 1269–1272.
OPEC. World oil outlook. Outlook. Vienna: Organisation of Petroleum
Exporting Countries, 2011.
Órfão JJM, Martins FG. Kinetic analysis of thermogravimetric
data obtained under linear temperature programming – a
method based on calculations of the temperature integral by
interpolation. Thermochim Acta 2002; 390: 195–211.
Ozawa T. A new method of analyzing thermogravimetric data. Bull
Chem Soc Japan 1965; 38: 1881–1886.
Pepper AS, Corvi J. Simple kinetic models of petroleum formation.
Part 1: oil and gas generation from kerogen. Mar Petrol Geol
1995; 12: 291–319.
Peters KE, Walters CC, Mankiewicz J. Evaluation of kinetic
uncertainty in numerical models of petroleum generation.
AAPG Bull 2006; 90: 387–403.
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM

Peters KE, Burnham AK, Walters CC. Petroleum generation kinetics:
single versus multiple heating-ramp open-system pyrolysis.
AAPG Bull 2015; 99: 591–616.
Qing W, Feng X, Baizhong S, Jingru B, Shaohua L. Study on the
kinetic characteristics in pyrolysis process of Huadian oil
shales by isoconversional method. Challenges of Power
Engineering and Environment. Berlin Heidelberg: Springer,
Part: 14, 2007: 1269–1274.
Rajeshwar K. The kinetics of the thermal decomposition of green
river oil shale kerogen by non-isothermal thermogravimetry.
Thermochim Acta 1981; 45: 253–263.
Rajeshwar K, Nottenburg R, Dubow J. Thermophysical properties of
oil shales. J Mater Sci 1979; 14: 2025–2052.
Reynolds JG, Burnham AK, Mitchell TO. Kinetic-analysis of California
petroleum source rocks by programmed-temperature
micropyrolysis. Org Geochem 1995; 23: 109–120.
Rick SJJ, Meilani P, Jim S. Advances in steady state process modeling
of oil shale retorting. 29th Oil Shale Symposium, Colorado
School of Mines, 2009.
Ronghua G, Lijun J, Jianbo Z, Jianyou W, Haoquan H. Effect of
inherent and additional pyrite on the pyrolysis behavior of oil
shale. J Anal Appl Pyrol 2014; 105: 342–347.
Schaefer RG, Schenk HJ, Hardelauf H, Harms R. Determination of
gross kinetic parameters for petroleum formation from Jurassic
source rocks of different maturity levels by means of laboratory
experiments. Org Geochem 1990; 16: 115–120.
Schenk HJ, Horsfield B. Kinetics of petroleum generation by
programmed-temperature closed-versus open-system
pyrolysis. Geochim Cosmochim Acta 1993; 57: 623–630.
Shabbar S, Qudaih R, Talab I, Janajreh I. Kinetics of pyrolysis and
combustion of oil shale sample from thermogravimetric data.
Fuel 2011; 90: 1631–1637.
Shuyuan L, Changtao Y. Study of pyrolysis kinetics of oil shale. Fuel
2003; 82: 337–342.
Shuyuan L, Changtao Y. Study of different kinetic models for oil
shale pyrolysis. Fuel Process Technol 2004; 85: 51–61.
Siskin M, Katritzky AR. Aqueous organic chemistry: Geochemical
aspects. Composition, Geochemistry and Conversion of Oil
Shales. Netherlands: Springer, 1995: 313–327.
Skala D, Sokic M. The determination of a complex kinetic expression
of oil shale pyrolysis using combined non-isothermal and
isothermal TG. J Therm Anal 1992; 38: 729–738.
Skala D, Kopsch H, Sokić M, Neumann HJ, Jovanović J.
Thermogravimetrically and differential scanning calorimetrically
derived kinetics of oil shale pyrolysis. Fuel 1987; 66: 1185–1191.
Skala D, Sokic M, Kopsch H. Oil shale pyrolysis – a new approach
to the kinetic investigation of different kerogen type samples.
Thermochim Acta 1988; 134: 353–358.
Skala D, Kopsen H, Sokić M, Neumann HJ, Jovanović J. Modelling and
simulation of oil shale pyrolysis. Fuel 1989; 68: 168–173.
Stankiewicz BA, Kruge MA. Flash Pyrolysi-GC-MS investigation
of maceral concentrates separated from Miocene Monterey
formation kerogen and Eocene Indonesian coal. Paper
presented to Division of Geochemistry, 207th ACS National
Meeting, San Diego, 1994.
Strizhakova YA, Usova TV. Current trends in the pyrolysis of oil shale:
a review. Solid Fuel Chem 2008; 42: 197–201.
Sun Y, Bai F, Liu B, Liu Y, Guo M, Guo W, Wang Q, Lü X, Yang F,
Yang Y. Characterization of the oil shale products derived via
topochemical reaction method. Fuel 2014; 115: 338–346.
M.A. Raja et al.: Oil shale pyrolysis and kinetics      21
Sun Y, Bai F, Lü X, Jia C, Wang Q, Guo M, Li Q, Guo W. Kinetic study
of Huadian oil shale combustion using a multi-stage parallel
reaction model. Energy 2015a; 82: 705–713.
Sun YH, Bai FT, Lü XS, Li Q, Liu YM, Guo MY, Guo W, Liu BC. A novel
energy-efficient pyrolysis process: self-pyrolysis of oil shale
triggered by topochemical heat in a horizontal fixed bed. Sci
Rep 2015b; 5: 8290.
Sundararaman P, Teerman SC, Mann RG, Mertani B. Activation
energy distribution: a key parameter in basin modelling and
a geochemical technique for studying maturation and organic
facies. In: Proceedings, Indonesian Petroleum Association,
17th Annual Convention, 1988: 169–185.
Tegelaar EW, Noble RA. Kinetics of hydrocarbon generation as a
function of the molecular structure of kerogen as revealed by
pyrolysis-gas chromatography. Org Geochem 1994; 22: 543–574.
Tissot B, Welte DH. Petroleum formation and occurrence, 2nd ed.
Berlin: Springer-Verlag, 1984.
Tiwari P, Deo M. Compositional and kinetic analysis of oil shale
pyrolysis using TGA-MS. Fuel 2012a; 94: 333–341.
Tiwari P, Deo M. Detailed kinetic analysis of oil shale pyrolysis TGA
data. AIChE J 2012b; 58: 505–515.
Torrente MC, Galán MA. Kinetics of the thermal decomposition
of oil shale from Puertollano (Spain). Fuel 2001; 80: 327–334.
Vahur O, Eric MS. Oil shale processing, chemistry and
technology. In: Fossil energy: selected entries from the
encyclopedia of sustainability science and technology.
New York: Springer, 2013: 99–148.
Vandenbroucke M. Kerogen: from types to models of chemical
structure. Oil Gas Sci Technol 2003; 58: 243–269.
Vandenbroucke M, Largeau C. Kerogen origin, evolution and
structure. Org Geochem 2007; 38: 719–833.
Vyazovkin SV, Lesnikovich AL. Practical application of isoconversional
methods. Thermochim Acta 1992; 203: 177–185.
Vyazovkin S, Burnham AK, Criado JM, Perez MLA, Popescu C,
Sbirrazzuoli N. ICTAC recommendations for performing kinetic
computations on thermal analysis data. Thermochim Acta 2011;
520: 1–19.
Vyazovkin S, Chrissafis K, Di Lorenzo ML, Koga N, PIjolat M, Roduit
B, Sbrirrazzuoli N. ICTAC kinetics committee recommendations
for collecting thermal analysis data for kinetic computations.
Thermochim Acta 2014; 590: 1–23.
Wallman PH, Tamm PW, Spars BG. Oil shale retorting kinetics. In:
Stauffer HC, editor, Oil shale, tar sands and related materials.
ACS Symposium Series 163. Washington, DC: American
Chemical Society, 1981.
Wanjun T, Yuwen L, Hen Z, Cunxin W. New approximate formula for
Arrhenius temperature integral. Thermochim Acta 2003; 408:
39–43.
Williams PT, Ahmad N. Investigation of oil-shale pyrolysis
processing conditions using thermogravimetric analysis. Appl
Energ 2000; 66: 113–133.
Xiangxin H, Kulaots I, Jiang X, Suuberg EM. Review of oil shale semicoke
and its combustion utilization. Fuel 2014; 126: 143–161.
Xiaoping L, Jin HZ, Dengguo L, Xiaoxing L, Zhanjun Z, Guangwen
X. Initial pyrolysis mechanism of oil shale kerogen with
reactive molecular dynamics simulation. Energ Fuel 2015; 29:
2987–2997.
Xiaoshu L, Youhong S, Tao L, Fengtian B, Martti V. An efficient and
general analytical approach to modelling pyrolysis kinetics of
oil shale. Fuel 2014; 135: 182–187.
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM
22      M.A. Raja et al.: Oil shale pyrolysis and kinetics
Yang Q, Qian Y, Andrzej K, Huairong Z, Yang S. Advanced energy
analysis of an oil shale retorting process. Appl Energ 2016a;
165: 405–415.
Yang Q, Qian Y, Andrzej K, Huairong Z, Yang S. Framework for
advanced exergoeconomic performance analysis and
optimization of an oil shale retorting process. Energy 2016b;
109: 62–76.
Youhong S, Fengtian B, Baochang L, Yumin L, Mingyi G, Wei G,
Qiuwen W, Xiaoshu L, Fang Y. Characterization of the oil shale
products derived via topochemical reaction method. Fuel 2014;
115: 338–346.
Youtsos MSK, Mastorakos E, Cant RS. Numerical simulation of
thermal and reaction fronts for oil shale upgrading. Chem Eng
Sci 2013; 94: 200–213.
Ziegel ER, Gorman JW. Kinetic modelling with multiresponse data.
Technometrics 1980; 22: 139–151.
Bionotes
Muhammad Afzal Raja
Beijing Key Laboratory of Ionic Liquids Clean
Process, Institute of Process Engineering,
Chinese Academy of Sciences, Beijing
100049, China; and College of Chemical and
Engineering, University of Chinese Academy
of Sciences, Beijing 100049, China
Muhammad Afzal Raja is currently completing his doctoral
research at the Institute of Process Engineering, University of
Chinese Academy of Sciences. His research areas include mod-
eling oil shale pyrolysis and kinetics, process simulation and
assessment of hydrogen and high-value hydrocarbon fuel products
from oil shale pyrolysis. He served at Pakistan Oilfields Ltd. (May
2011–August 2013) and in Descon Integrated Projects Ltd. (August
2008–April 2013) as a process safety engineer. He received his
graduation degree in 2007 from the University of Engineering and
Technology, Lahore, Pakistan, and received his Master’s degree in
2009 from the University of Punjab, Pakistan. He has two presenta-
tions to his credit.
Yongsheng Zhao
Beijing Key Laboratory of Ionic Liquids Clean
Process, Institute of Process Engineering,
Chinese Academy of Sciences, Beijing
100049, China; and College of Chemical and
Engineering, University of Chinese Academy
of Sciences, Beijing 100049, China
Yongsheng Zhao is currently completing his doctoral research at the
Institute of Process Engineering, University of Chinese Academy of
Sciences, and he will graduate in June 2017. His major is chemical
engineering and technology. His research areas include ionic liquids
for various gases absorption and separation: an extensive database
with rich data and a systemic screening method. He has published
seven SCI papers and one conference paper.
Xiangping Zhang
Beijing Key Laboratory of Ionic Liquids Clean
Process, Institute of Process Engineering,
Chinese Academy of Sciences, Beijing
100049, China
xpzhang@ipe.ac.cn
Xiangping Zhang received her PhD degree in Chemical Engineering
in 2002 from the Dalian University of Technology. Then she joined
Institute of Process Engineering of CAS. Her research interests
include process integration, thermodynamics and separation
engineering. She has published about 130 papers in peer-reviewed
journals, and obtained 20 issued Chinese invention patents and 4
issued PCT patents. She was awarded the National Award for Natural
Sciences and the Nomination Award of the 4th Top Ten Outstanding
Women in Chinese Academy of Sciences in 2012.
Chunshan Li
Beijing Key Laboratory of Ionic Liquids Clean
Process, Institute of Process Engineering,
Chinese Academy of Sciences, Beijing
100049, China
Chunshan Li received his PhD degree in Chemical Engineering from
the Institute of Process Engineering in 2006. He was a postdoctoral
fellow at Federal Institute for Materials Research and Testing, BAM
(2006–2007), and a JSPS fellow at Nagoya University, Japan (2007–
2009). He joined IPE, CAS and has been a professor at the Hundred
Talents Project of IPE since September 2009. His research mainly
focuses on a cleaner energy catalytic process, chemical synthesis,
chemical process design and integration. He has published more
than 70 peer-reviewed articles in journals such as AIChE J., Ind. Eng.
Chem. Res. and Chem. Eng. J. and holds 29 patents.
Suojiang Zhang
Beijing Key Laboratory of Ionic Liquids Clean
Process, Institute of Process Engineering,
Chinese Academy of Sciences, Beijing
100049, China
Suojiang Zhang is Professor and Director General of the Institute of
Process Engineering (IPE), Chinese Academy of Sciences (CAS), a
member of CAS, Director of Beijing Key Laboratory of Ionic Liquids
Clean Process, and Director of Professional Committee of Chinese
Chemical Society of Ionic Liquids and Green Engineering. He is
“National Science Fund for Distinguished Young Scientists” Winner,
973 Chief Scientist, and fellow of the Royal Society of Chemistry. He
mainly engages in ionic liquids and green process engineering. He has
published more than 300 papers and more than 90 patents, and also
edited or written four monographs. He won a number of awards, such
as the Second Class Prize of National Natural Science Award in 2010.
Brought to you by | University of Colorado - Boulder
Authenticated
Download Date | 4/1/17 10:02 AM